Figure 6.

YOLO-based stomatal detection training and its evaluation. (A) Preparation of training dataset using manual labeling and the resulting prediction by trained model. (B) The statistical evaluation of the trained model by precision and recall metric. (C) Stomatal labeled images based on our train YOLO model. (D) Model evaluation using validation set during the training based on the metrics including precision (upper left), recall (upper right), mAP_0.5 (bottom left), mAP_0.5:0.95 (bottom right). (E) Precision and recall evaluation on a test set with a trained model.